PAGANELLI MATTEO

Entity Matching and data integration - Development of Machine Learning and Deep Learning techniques to automate Entity Matching and data integration, with a focus on robustness in scenarios characterized by information scarcity and noisy data. The research includes approaches primarily based on deep learning models and explainability methods to ensure transparency and trust in the results.

Explainability and Counterfactuals - Research on model interpretability techniques, including counterfactual explanations, aimed at improving the understanding of automated decisions and fostering the responsible use of AI systems in complex and sensitive scenarios.

LLMs for Textual and Tabular Data - Research on the use of Large Language Models for the integrated analysis of textual and tabular data, with a focus on table-based question answering (both single-table and multi-table) applied to complex reports (particularly non-financial ones). Applications include information extraction, fact-checking, and greenwashing detection, supported by dedicated benchmarks.